mcpField
G
G2N MCP Google Calendar SSE Server

G2N MCP Google Calendar SSE Server

Remote
github.com/gabriel-g2nProductivity7Updated 82w ago
Repository

About G2N MCP Google Calendar SSE Server

An SSE-based MCP server that exposes Google Calendar management tools to compatible models and applications. It supports calendar and event lookup, creation, updates, and deletion, with OAuth-based access.

Key features

  1. 01List calendars and events
  2. 02Read calendar and event details
  3. 03Create, update, and delete events
  4. 04List calendar and event colors
  5. 05OAuth token refresh

Use cases

  1. 01Reviewing events through an MCP client
  2. 02Creating or updating calendar entries
  3. 03Connecting calendar tools to n8n workflows

MCP support

The server exposes Google Calendar operations as MCP tools over SSE for compatible clients and applications.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttp://localhost:3001/sse
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

G2N MCP Google Calendar SSE Server

🌎 This README is available in multiple languages:

Overview

The G2N MCP Google Calendar SSE Server is a Model Context Protocol (MCP) server implementation that provides Google Calendar integration through Server-Sent Events (SSE). This server exposes Google Calendar functionality as tools that can be used by AI models and applications like Cursor, Claude, and n8n to interact with Google Calendar.

Built with the latest MCP SDK version, this server provides robust integration between MCP-compatible models and Google Calendar services.

Features

The server provides the following MCP tools for Google Calendar management:

  • list-calendars: List all available calendars
  • get-calendar: Get details of a specific calendar
  • list-events: List events from a calendar with filtering options
  • get-event: Get detailed information about a specific event
  • create-event: Create a new calendar event
  • update-event: Update an existing calendar event
  • delete-event: Delete a calendar event
  • list-colors: List available colors for events and calendars
What's New in v1.1.0
  • Comprehensive logging system with configurable log levels
  • Request/response logging middleware for better monitoring
  • SSE heartbeat implementation for stable connections
  • Enhanced error handling and debugging capabilities
  • Improved Docker configuration with proper data directory permissions
  • Enhanced OAuth flow with better token management
  • Code refactoring for better maintainability
What's New in v1.0.1
  • First stable release
  • Production-ready with Docker and Docker Swarm support
  • Enhanced environment variables configuration
  • Improved connection URLs for different deployment scenarios
  • Better documentation for n8n integration
  • Traefik configuration guidelines
  • Multi-platform Docker images (amd64, arm64, arm/v7)

Getting Started

Prerequisites
  • Docker and Docker Compose installed
  • Google Cloud project with Calendar API enabled
  • OAuth 2.0 Client ID and Client Secret
Environment Variables

The server uses the following environment variables:

PORT=3001                                # Server port (default: 3001)
PUBLIC_URL=https://your-domain.com       # Public URL for OAuth callbacks
GOOGLE_CLIENT_ID=${GOOGLE_CLIENT_ID}     # Google OAuth Client ID
GOOGLE_CLIENT_SECRET=${GOOGLE_CLIENT_SECRET} # Google OAuth Client Secret
OAUTH_REDIRECT_PATH=/oauth/callback      # OAuth callback path (default: /oauth/callback)

Important Notes:

  • When using Traefik, make sure to configure it to point to the port specified in the PORT environment variable
  • This is crucial for successfully receiving the Google refresh token
  • The PUBLIC_URL must be internet-accessible for OAuth callbacks to work

Authentication Flow

  1. Start the server using Docker or directly
  2. Navigate to the /auth endpoint in your browser
  3. Grant permissions to the application using your Google account
  4. After authorization, the server will store refresh tokens for continuous access
  5. The server will automatically refresh tokens when needed

To revoke access, use the /revoke endpoint:

curl -X POST https://your-domain.com/revoke

Usage with MCP-Compatible Applications

Connection URLs

Depending on your deployment scenario, use the appropriate URL format:

  1. Docker Swarm / n8n internal access:

    http://[stack-service-name]:3001/sse
    

    Example: If your service is named g2n-mcp-gcal-sse in the stack, use:

    http://g2n-mcp-gcal-sse:3001/sse
    
  2. External access (Cursor, Claude, etc.):

    https://your-domain.com/sse
    
  3. Local development:

    http://localhost:3001/sse
    
Cursor AI

You can use this server with Cursor AI by configuring the MCP connection in your settings:

  1. Open Cursor settings
  2. Configure the MCP server URL using your public domain:
    https://your-domain.com/sse
    
  3. Start using Google Calendar features through AI commands
Claude Desktop

For Claude Desktop:

  1. Navigate to Settings > MCP
  2. Add a new MCP connection with your public URL:
    https://your-domain.com/sse
    
  3. Access Google Calendar functionality through your conversations
n8n
  1. In n8n, add a new MCP node
  2. Configure the MCP node with the internal service URL:
    http://[stack-service-name]:3001/sse
    
  3. Use the exposed calendar tools in your workflows

License

This project is licensed under the MIT License - see the LICENSE file for details.

About G2NTech

This project is developed and maintained by Gabriel Augusto at G2NTech.

Support the Project 💜

If you find this project useful, consider supporting it via PIX:

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.